Identification and validation of autophagy-related genes in keratoconus and their correlation with immune...

Keratoconus (KCN) is a bilateral and asymmetric cornea disease. Autophagy plays an important role in homeostasis by protecting cells against stress. However, the roles of autophagy-related genes (ARGs) in KCN remains unclear. Hence, this study aimed to identify the signatures of ARGs of KCN and explore their correlation with immune infiltration. Transcriptional data and clinical information of patients with KCN were downloaded from the profile data GSE112155 and GSE151631. Functional analysis was used to reflect the biological functions, and weighted gene co-expression network analysis was applied to excavate co-expression modules of autophagy-related expression patterns. Moreover, gene set enrichment and variation analyses were performed for pathway analysis. Consensus clustering analysis was used to cluster different molecular subtypes on the basis of gene expression profiles of KCN-specific ARGs. Single-sample gene set enrichment analysis was employed to calculate separate enrichment scores for each immunocyte between KCN and healthy samples. Finally, hub genes were verified by real-time quantitative polymerase chain reaction. We first identified 14 ARGs differentially expressed between patients with KCN and controls using NetworkAnalyst. Nine overlapped genes (BNIP3, CDKN1A, DDIT3, FOS, HSPA5, MAPK8IP1, MYC, PPP1R15A, and VEGFA) (P < .05) were identified using a random forest model. The MAPK signaling pathway, apoptosis, FoxO signaling pathway, and protein processing in endoplasmic reticulum signaling were mainly involved. The weighted gene co-expression network analysis classified the genes into 12 distinct modules. The MEturquoise (correlation = 0.51), MEpink (correlation = 0.535) were significantly positively correlated with KCN, whereas the MEyellow (correlation = -0.776) and MEgreen (correlation = -0.664) were negatively correlated. Single-sample gene set enrichment analysis showed a close interaction between immune cell infiltration and the development of KCN. Finally, all the 9 hub genes except VEGFA were significantly downregulated (P < .05) using real-time quantitative polymerase chain reaction. We described the signatures of ARGs in KCN, the distribution of immune cells between KCN patients and the control, and the correlation between hub genes related to autophagy and KCN disease. This demonstrates the potential roles of autophagy mechanisms and the immune response in KCN, providing a novel insight into understanding the pathogenesis of KCN and potential treatment targets.